*** kevinz has joined #openstack-nimble | 00:56 | |
*** Jeffrey4l has joined #openstack-nimble | 01:23 | |
openstackgerrit | liusheng proposed openstack/python-nimbleclient: Support availability zone in nimbleclient https://review.openstack.org/411690 | 01:43 |
---|---|---|
liusheng | zhenguo: ^^^ | 01:44 |
liusheng | morning nimble! | 01:44 |
*** l4yerffeJ_ has quit IRC | 01:44 | |
*** l4yerffeJ_ has joined #openstack-nimble | 01:44 | |
zhenguo | liusheng: o/ | 01:54 |
shaohe_feng | morning | 01:54 |
shaohe_feng | zhenguo: Where do we remove the retry node when rescheduler? | 01:55 |
zhenguo | shaohe_feng: you mean the node failed to deploy? | 01:56 |
shaohe_feng | zhenguo: yes. | 01:57 |
shaohe_feng | zhenguo: for example. get 10 nodes [node0, .... , node9] candidate. | 01:57 |
shaohe_feng | zhenguo: first scheduler to node0, failed. | 01:58 |
shaohe_feng | zhenguo: then, we need to remove the node0, and try node1 | 01:58 |
zhenguo | shaohe_feng: currently I remove the node from cache after schedule, but If the cache refresh again and the ironic node doesn't go to a unavailable state, it may have some problems | 01:59 |
shaohe_feng | zhenguo: I find node0 will be in retry["nodes"] | 01:59 |
zhenguo | shaohe_feng: yes, it's in scheduler filter_properties | 02:00 |
zhenguo | shaohe_feng: maybe we can add a check in scheduler, to ignore node which in retry[] | 02:01 |
shaohe_feng | zhenguo: Then why not set(candidate) - set(retry["nodes"]) | 02:02 |
zhenguo | shaohe_feng: good suggestion | 02:03 |
zhenguo | shaohe_feng: do you retired the node again on your test? | 02:05 |
shaohe_feng | zhenguo: retry failed last night. maybe ironic problem. | 02:06 |
shaohe_feng | zhenguo: I'm trying again. | 02:07 |
zhenguo | shaohe_feng: I have merged the patch which adds lock when accessing node_cache, and remove the selected node from cache, please try with that | 02:08 |
openstackgerrit | liusheng proposed openstack/python-nimbleclient: Support availability zone in nimbleclient https://review.openstack.org/411690 | 02:08 |
shaohe_feng | zhenguo: OK. with the new patch. | 02:10 |
*** l4yerffeJ__ has joined #openstack-nimble | 02:14 | |
*** l4yerffeJ_ has quit IRC | 02:14 | |
openstackgerrit | liusheng proposed openstack/python-nimbleclient: Support availability zone in nimbleclient https://review.openstack.org/411690 | 02:15 |
*** Jeffrey4l has quit IRC | 02:22 | |
openstackgerrit | liusheng proposed openstack/python-nimbleclient: Remove python2.4 compatibility support https://review.openstack.org/412697 | 02:22 |
openstackgerrit | liusheng proposed openstack/python-nimbleclient: Add Python 3.5 classifier and venv https://review.openstack.org/412700 | 02:27 |
*** yuntongjin has joined #openstack-nimble | 02:40 | |
*** l4yerffeJ_ has joined #openstack-nimble | 02:52 | |
shaohe_feng | zhenguo:$ git pull │ | 02:53 |
shaohe_feng | fatal: Could not read from remote repository. │ | 02:53 |
shaohe_feng | │ | 02:53 |
shaohe_feng | Please make sure you have the correct access rights │ | 02:53 |
shaohe_feng | and the repository exists. | 02:53 |
*** l4yerffeJ__ has quit IRC | 02:55 | |
liusheng | shaohe_feng: maybe just retry | 02:56 |
zhenguo | shaohe_feng, liusheng: sorry, I am in a internal meeting, will be back soon | 02:57 |
* zhenguo meeting | 02:57 | |
shaohe_feng | zhenguo: liusheng: I'm also in an internal meeting. :) | 03:01 |
liusheng | shaohe_feng: :) | 03:03 |
*** Jeffrey4l has joined #openstack-nimble | 03:14 | |
zhenguo | shaohe_feng, liusheng: I'm back | 03:16 |
zhenguo | shaohe_feng, liusheng: but maybe I have to support a POC next few days :( | 03:17 |
liusheng | zhenguo: :(, will you have a business trip ? | 03:21 |
zhenguo | liusheng: no, just support in office | 03:21 |
liusheng | zhenguo: that will spend your much time | 03:23 |
zhenguo | liusheng: I think it is, yes | 03:24 |
liusheng | zhenguo: it is a pity, PTL will AFK :) | 03:26 |
zhenguo | liusheng: lol, | 03:28 |
zhenguo | liusheng: I will try to keep online, | 03:28 |
liusheng | zhenguo: I will wait to captrue to review, hah | 03:30 |
*** l4yerffeJ_ has quit IRC | 03:46 | |
*** l4yerffeJ_ has joined #openstack-nimble | 03:46 | |
*** yuntongjin has quit IRC | 04:39 | |
*** yuntongjin has joined #openstack-nimble | 05:11 | |
*** yuntongjin has quit IRC | 05:41 | |
shaohe_feng | liusheng: zhenguo: OK | 06:56 |
shaohe_feng | ping | 06:56 |
shaohe_feng | do nimble support availability_zone? | 06:56 |
liusheng | shaohe_feng: yes | 06:56 |
zhenguo | shaohe_feng: yes | 06:56 |
shaohe_feng | OK. an new API? | 06:57 |
shaohe_feng | how do we create a availability_zone | 06:57 |
shaohe_feng | by a API? | 06:57 |
shaohe_feng | liusheng: zhenguo: I'm write a summary to JF. :) | 06:57 |
zhenguo | shaohe_feng: we get az from ironic properties | 06:58 |
shaohe_feng | zhenguo: How do we set the az? | 06:58 |
shaohe_feng | zhenguo: we need to set az for different nodes. | 06:58 |
zhenguo | shaohe_feng: set availability_zone='xxx' to ironic node properties like instance type | 06:59 |
shaohe_feng | zhenguo: liusheng: nimble support 3 filters: availability_zone, capabilities_filter, instance_type | 06:59 |
liusheng | zhenguo, shaohe_feng may you can also help to review the as support in client side: https://review.openstack.org/#/c/411690/ :) | 06:59 |
liusheng | s/as/az/ | 06:59 |
shaohe_feng | zhenguo: liusheng: I do not see concrete weight in nimble. | 07:00 |
shaohe_feng | liusheng: OK. will do it after the summary. | 07:01 |
shaohe_feng | liusheng: good. | 07:01 |
shaohe_feng | liusheng: it is availability zone support. | 07:01 |
liusheng | zhenguo: yes | 07:02 |
liusheng | shaohe_feng: yes | 07:02 |
zhenguo | liusheng: ok, will make sure them be merged in by today :) | 07:02 |
zhenguo | shaohe_feng: yes, no weight currently, and seems not needed | 07:02 |
liusheng | zhenguo: thank you sir! lol | 07:02 |
zhenguo | liusheng: hah | 07:02 |
zhenguo | shaohe_feng: maybe need weight to select nodes with different ports, like one has 2 ports and others has more than 2 ports | 07:03 |
*** yuntongjin has joined #openstack-nimble | 07:04 | |
* zhenguo away | 07:05 | |
shaohe_feng | zhenguo: yes. maybe we also need to filter for ports | 07:05 |
shaohe_feng | liusheng: multi-ports filter maybe need. | 07:06 |
liusheng | shaohe_feng: good idea, that maybe future task | 07:07 |
shaohe_feng | liusheng: yes. In plan. | 07:07 |
shaohe_feng | liusheng: so the tempest tests has support, right? | 07:13 |
liusheng | shaohe_feng: it is still be stucked | 07:14 |
liusheng | shaohe_feng: I mean the instance creation tempest | 07:14 |
liusheng | shaohe_feng: the instance_tpye tests is ok | 07:14 |
shaohe_feng | liusheng: Got it. Thanks. | 07:15 |
shaohe_feng | zhenguo: liusheng: $ vim doc/source/api.rst | 07:23 |
shaohe_feng | I only get the instance related API description | 07:24 |
shaohe_feng | on instance-type. | 07:24 |
shaohe_feng | no instance-type | 07:24 |
liusheng | shaohe_feng: seems it is missing | 07:26 |
shaohe_feng | liusheng: zhenguo: http://paste.openstack.org/show/592868/ | 07:40 |
shaohe_feng | about the summary | 07:40 |
shaohe_feng | any supplement? | 07:41 |
shaohe_feng | enrich? | 07:41 |
liusheng | actions of instance.(create, delete, list, show) -- also support update | 07:43 |
liusheng | shaohe_feng: and az support in api | 07:43 |
shaohe_feng | liusheng: Thanks. also update the https://etherpad.openstack.org/p/nimble-task | 07:47 |
shaohe_feng | liusheng: we need to update the docs. and this link can help remider us to do it. | 07:47 |
liusheng | shaohe_feng: ok, I will pick the task if I have enough time :) | 07:48 |
openstackgerrit | liusheng proposed openstack/nimble: Use oslo context instead of the native context implementation https://review.openstack.org/412350 | 07:55 |
openstackgerrit | liusheng proposed openstack/nimble: Remove unused print statement https://review.openstack.org/412852 | 08:01 |
shaohe_feng | liusheng: for the context, it maintains a session for sqlalchemy | 08:09 |
liusheng | shaohe_feng: sorry, what's that mean | 08:10 |
shaohe_feng | liusheng: https://github.com/openstack/nimble/blob/master/nimble/db/sqlalchemy/api.py#L69 | 08:11 |
shaohe_feng | liusheng: can we use the session. | 08:11 |
shaohe_feng | liusheng: I go through nova's code, find nova's ontext has a session field. | 08:11 |
shaohe_feng | liusheng: Guess oslo context also has a session. | 08:12 |
shaohe_feng | liusheng: let me show the nova code. | 08:12 |
liusheng | shaohe_feng: there isn't a session in oslo_context | 08:14 |
shaohe_feng | liusheng: https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api.py#L571 | 08:15 |
shaohe_feng | liusheng: got it. nova store the session in context. | 08:15 |
shaohe_feng | liusheng: I do not know the benefit. | 08:16 |
liusheng | shaohe_feng: I am not sure they are same things, but seems nimble don't use that | 08:18 |
liusheng | shaohe_feng: is the "context" in Nova is RequestContext object? | 08:18 |
shaohe_feng | liusheng: yes, it is. | 08:22 |
shaohe_feng | liusheng: https://github.com/openstack/nova/blob/master/nova/context.py#L64 | 08:22 |
shaohe_feng | zhenguo: liusheng: ^ | 08:22 |
shaohe_feng | this code, it use @enginefacade.transaction_context_provider to fill the session in RequestContext | 08:23 |
liusheng | shaohe_feng: hmm, seems it is. not sure is it benifit for Nimble | 08:24 |
shaohe_feng | liusheng: not sure, the session can be across the components in nova. | 08:27 |
liusheng | shaohe_feng: may we can improve that if we really think it is valuable in the future, for now, if not necessary, not by entity, what do you think ? lol | 08:30 |
shaohe_feng | liusheng: yes, agree. We have more important things to do. | 08:32 |
liusheng | shaohe_feng: agree! | 08:33 |
openstackgerrit | liusheng proposed openstack/nimble: Use oslo context instead of the native context implementation https://review.openstack.org/412350 | 08:46 |
*** kevinz has quit IRC | 09:05 | |
*** kevinz has joined #openstack-nimble | 09:05 | |
liusheng | shaohe_feng, zhenguo how about just remove nimble/doc/source/api.rst | 09:23 |
shaohe_feng | liusheng: sames it is OK. | 09:25 |
liusheng | shaohe_feng, zhenguo since we have a detailed api-ref | 09:25 |
shaohe_feng | liusheng: yes. | 09:29 |
zhenguo | liusheng, shaohe_feng: it's just a design API instruction, not all API there are implemented | 09:30 |
liusheng | zhenguo: you mean the api.rst, but it only include the instance APIs, and every api description only includes a method and url. the api-ref is more detailed | 09:32 |
liusheng | zhenguo: we also have api-sampels | 09:32 |
liusheng | zhenguo, shaohe_feng can you take a look at https://review.openstack.org/#/c/412350/, since it is easy to conflict with others changes, lol | 09:33 |
zhenguo | liusheng: yes, it's just a list of API that we want to implement, some of them are not exist now | 09:35 |
zhenguo | liusheng: we should remove it after all it's implemented | 09:35 |
liusheng | zhenguo: but the doc doesn't include instnace type related, az related APIs | 09:36 |
liusheng | zhenguo: it doesn't like a api work todo list | 09:36 |
zhenguo | liusheng: yes, hah, and you can remove the API which already exist | 09:36 |
liusheng | zhenguo: do yo prefer to supplyign than deletion ? | 09:37 |
zhenguo | liusheng: I want to have a file to list all non-implemented APIs | 09:37 |
liusheng | zhenguo: if for this reason, this file need to be delted, hah | 09:38 |
liusheng | zhenguo: it is a subset of the implemented APIs | 09:38 |
liusheng | zhenguo: and may not be this name "api.rst" | 09:38 |
zhenguo | liusheng: hah | 09:38 |
liusheng | zhenguo: lol | 09:38 |
zhenguo | liusheng: yes, it seems some kind of strange in our repo, | 09:39 |
zhenguo | liusheng: maybe you can just delete them and create a etherpad to list APIs we want to implement in future | 09:39 |
liusheng | zhenguo: sure, or just add a *api* section in our existed etherpad or wiki :) | 09:42 |
zhenguo | liusheng: ok, | 09:43 |
zhenguo | liusheng, shaohe_feng: wrt https://review.openstack.org/#/c/412852/ , does the log print always or only when there are some errors happend with the tests? | 09:46 |
liusheng | zhenguo: will always print | 09:47 |
shaohe_feng | zhenguo: $ git show b2b241f085d208450614d6a7b4fcdeabb2a610ad | 09:49 |
shaohe_feng | zhenguo: you add it. The commit does not tell us the reason you add it? | 09:50 |
*** kevinz has quit IRC | 09:56 | |
shaohe_feng | zhenguo: liusheng: go home | 10:15 |
zhenguo | liusheng: in fact, that LOG copied from ironic, but I'm OK with removing it, we can add it manually if some errors happened | 10:39 |
liusheng | zhenguo: ok | 10:39 |
zhenguo | shaohe_feng: yes, but seems 'Add update instance support' is enough :P | 10:40 |
*** yuntongjin has quit IRC | 10:45 | |
openstackgerrit | Zhenguo Niu proposed openstack/python-nimbleclient: Add support for setting power state for instance https://review.openstack.org/411729 | 11:22 |
openstackgerrit | Merged openstack/nimble: Remove unused print statement https://review.openstack.org/412852 | 11:24 |
openstackgerrit | Merged openstack/python-nimbleclient: Add support for updating instance https://review.openstack.org/411591 | 11:24 |
openstackgerrit | Merged openstack/python-nimbleclient: Support availability zone in nimbleclient https://review.openstack.org/411690 | 11:25 |
openstackgerrit | Merged openstack/nimble: Remove valueless api.rst in doc/source https://review.openstack.org/412975 | 11:34 |
*** l4yerffeJ_ has quit IRC | 12:20 | |
*** l4yerffeJ_ has joined #openstack-nimble | 12:20 | |
openstackgerrit | liusheng proposed openstack/nimble: Fix the instance type extra related APIs https://review.openstack.org/413053 | 12:20 |
openstackgerrit | liusheng proposed openstack/nimble: Correct the api samples of type extra APIs https://review.openstack.org/413056 | 12:29 |
openstackgerrit | liusheng proposed openstack/nimble: Correct the api samples of type extra APIs https://review.openstack.org/413056 | 12:33 |
*** yuntongjin has joined #openstack-nimble | 12:34 | |
*** yuntongjin has quit IRC | 13:29 | |
*** l4yerffeJ has joined #openstack-nimble | 21:33 | |
*** l4yerffeJ_ has quit IRC | 21:34 | |
*** Jeffrey4l has quit IRC | 21:35 | |
*** Jeffrey4l has joined #openstack-nimble | 21:47 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!